logcollector

[cgzones]

• fix some compiler warnings
• fix some static analyzer warnings

[awiddersheim]

Hrm, looks like the tests failed here:

https://travis-ci.org/ossec/ossec-hids/jobs/37510615#L2778

Looks like this test is supposed to fail:

https://github.com/ossec/ossec-hids/blob/master/src/tests/test_os_net.c#L288

I'm not really familiar with travis-ci or the tests. This seems like a one off failure? Did just this test timeout or all the tests as a whole hit some type of timeout?

I'd merge but since the tests hit a failure I don't think all the tests ran after this one. Is there a way to kick off another build without committing?

[awiddersheim]

Researched a bit more. Here is my understanding. I'm not seeing a timeout being set in our set of tests (could have just missed it) so the default is being set which according to this documentation is four seconds:

http://check.sourceforge.net/doc/check_html/check_4.html#Test-Timeouts

The test against OS_GetHost() specifies a retry attempt of two:

https://github.com/ossec/ossec-hids/blob/master/src/tests/test_os_net.c#L288

Looking at the OS_GetHost() function this actually ends up being three attempts sleeping a second between each attempt so one can expect this test to take at least three seconds. I'm not sure how long a call to gethostbyname() might take but this cuts kind of close to that default timeout and lowering the attempts in the test might be in order.

Hopefully, my interpretation of all of this is makes sense but I'd love to learn more so please let me know.

[cgzones]

Thanks for your investigation @awiddersheim.
I rebased this branch, so let us see if it fails again.
About the unit test: i will think over it